144 ALASKAN GLACIER STUDIES of the cascading part of the glacier, which permits the falling of ice fragments from its front, thus adding a very effective cause for recession to the ordinary one of melting. A second important reason for more rapid recession in this stage is oftentimes the withdrawal of the main glacier from beneath the cascading glacier. This withdrawal of the ice results in the raising of the mean annual temperature in the immediate neighborhood, and consequently in an increase in ablation.1 This cause becomes particularly effective when by the withdrawal of the main glacier the waters of a fiord take its place. Perhaps this is the main reason why the Hanging Glacier on the north side of Nunatak Fiord no longer cascades out of its mountain valley. For these two reasons, after disconnection of a cascading glacier from the main glacier there is a period of rapid recession; but above a certain level melting must again become retarded and recession diminish because of the lower mean temperature at that elevation. Ultimately, if no change in the ice supply comes about, the time might arrive when a balance between ablation and supply would be reached and the end of the cascading glacier hold its position somewhere below the level of the valley lip. Such condition does not seem to be possible in the Yakutat Bay region excepting up those main valleys which are still occupied by ice. Throughout the region, wherever the main glacier has receded past the mouths of the hanging valleys, it is apparent that but a short time has been required for recession to destroy the cascading condition and to force the ice front back up into the hanging mountain valley. This fact is also illustrated in the Nunatak Fiord. Almost directly opposite the Cascading Glacier, on the north side of the Nunatak Fiord valley, there is a hanging valley of somewhat lower level than that occupied by the Cascading Glacier, and, therefore, presumably formerly occupied by a more vigorous glacier. The first positive knowledge in regard to this valley was obtained from Gilbert's photograph in 1899 (PI. LXTV) where it is seen that the tidal arm of Nunatak Glacier at that time expanded completely across the mouth of this hanging valley. In the valley was Hanging Glacier, an ice tongue in which clear ice showed just above the lip of the hanging valley, while below the lip a moraine-covered tongue projected to about half the distance between the lip and the surface of Nunatak Glacier. Because of the lowness of this hanging valley, and because of the fact that it opens toward the south, ablation had proceeded much faster on the end of this glacier than on the Cascading Glacier. It was, therefore, so covered by moraine that the appearance of cascading was entirely absent, a condition due partly to the fact that the slope below the lip of the hanging valley is less steep than that in the Cascading Glacier, partly to the presence of an extensive deposit of gravels on the northern side of Nunatak valley rising part way up the mountain base. In the interval between 1899 and 1909 Hanging Glacier receded so perceptibly that its lower moraine-covered end was almost up to the level of the hanging valley lip; and in the meantime the clear ice area was entirely replaced by moraine-covered ice. At the present rate of recession it cannot be long before this glacier will lie entirely inside the lip of the hanging valley. It may, therefore, be considered a third stage in the destruction of the cascading type of glacier. Succeeding stages introduce no points of special significance, for the later stages consist merely in continued recession of a valley glacier. In the valley next west of the Cascading Glacier, which hangs about 700 feet above the fiord level (PI. LI, B), there is such a receding glacier whose end now lies about a mile back from the lip of the hanging valley. In this case a single good-sized glacial stream flows from the ice front down to •*The covering of a main gkcier by a thick mantle of ablation moraine during recession has the same effect.